(this question were asked after studying line integrals)

1- Show that if $C$ is the graph of $y=f(x)$, $a \leq x \leq b,$ and if $F$ is a function of 2 variables defined on C, then $$\int_{C} F(x,y)dx = \int_{a}^{b} F(x, f(x))dx. $$

But I do not understand the meaning of the question, does it means that if I change the graph by its parametrization I will change the integration along the graph to the integration at the end points of the interval $[a,b]$, if so why is this correct?

Also could anyone give me a hint for the proof please?


  • $\begingroup$ Are you sure there is not a Jacobian missing, e.g. $\sqrt{1+f'(x)^2}$ ? $\endgroup$ – Thomas Apr 18 at 7:09
  • $\begingroup$ @Thomas no there is not ..... this is exactly the question I have to answer. $\endgroup$ – Intuition Apr 18 at 7:11
  • 1
    $\begingroup$ Looks a bit weird to me but maybe I am not understanding. The left member when F(x,y)=1 should not be the length of C ? When parametrizing the curve by $t \rightarrow [t,f(t)]$ tutorial.math.lamar.edu/Classes/CalcII/ParaArcLength.aspx we would get a Jacobian more ? Maybe I am not getting the meaning of the let side... $\endgroup$ – Thomas Apr 18 at 7:19
  • $\begingroup$ May be the question contains a typo (which I am not sure of it because my professor have seen the question and did not say anything about it ..... or may be he\she was not concentrating at that moment) ..... I do not know @Thomas $\endgroup$ – Intuition Apr 18 at 7:29
  • 1
    $\begingroup$ Maybe I got the catch. I did not notice that you have a dx in the left member, which must be maybe intended as the integral of the differential form "F dx" over C. Before I interpreted the left integral as simply the integration of a scalar function over the manyfold C. $\endgroup$ – Thomas Apr 18 at 12:37

$C$ is defined by $$C := \{ (x,y) \ | \ x\in[a,b], \ f(x) = y \}$$ so $$\int_{C} F(x,y)dx = \int_{\substack{x\in [a,b] \\ y = f(x)}} F(x,y)dx = \int_{x\in[a,b]} F(x,f(x))dx =\int_a^b F(x,f(x))dx$$ This just represents any function along $C$. For example the length of the curve would be $$\int_C \sqrt{dx^2 + dy^2} = \int_C\sqrt{1 + (\frac{dy}{dx})^2}dx = \int_a^b \sqrt{1 + (f'(x))^2}dx$$ Here $F(x,y) = \sqrt{1 + (\frac{dy}{dx})^2}$

  • $\begingroup$ what about the proof? $\endgroup$ – Intuition Apr 18 at 10:46
  • $\begingroup$ I made an edit which should be enough justification. It could almost be used as a definition in my opinion. $\endgroup$ – Dayton Apr 18 at 10:56
  • $\begingroup$ Thank you so much for you great effort. $\endgroup$ – Intuition Apr 18 at 10:59
  • $\begingroup$ why is the first integration on the right hand side of the second line from below does not contain $dx$? $\endgroup$ – Smart Apr 18 at 13:46
  • $\begingroup$ The infinitesimal Euclidean arc length is just $\sqrt{dx^2 + dy^2}$, which you can "factor" the $dx$ outside and have an integral depending only on one parameter. $\endgroup$ – Dayton Apr 18 at 13:56

Here one derivation starting from the definitions. Let $\tau:[a,b]\rightarrow R^2:t\rightarrow(t,f(t))$ a parametrization of $C$.

By definition the integral of a differential form $\rho$ is:

$\int_C \rho =\int_{[a,b]} \rho(\tau(t))[\tau'(t)]dt$

, where $\rho(a)[b]$ is the differential form at $a$ acting on vector $b$.

when $\rho=Fdx$, since $\tau'(t)=(1,f'(t))$, $dx[\tau'(t)]=1$ we have:

$\int_C F(x,y) dx =\int_{[a,b]} F(t,f(t))dt$

when $\rho=Fdy$, since $\tau'(t)=(1,f'(t))$, $dy[\tau'(t)]=f'(t)$ we have:

$\int_C F(x,y) dy =\int_{[a,b]} F(t,f(t))f'(t)dt$


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.